Abstract:Radio base stations are quite crucial to the quality of mobile communication services. The path loss of the mobile signal is mainly affected by the physical parameters of the base station itself and the geographical structure within the service area. Efficiently optimizing the number and assignment of radio base stations by predicting the path loss of radio signals within the service range is of great significance for improving signal quality and saving costs. The traditional modeling method based on the physical parameters of the base station and wireless signal propagation mechanism cannot consider geographical structure information, and the prediction accuracy is low. Another method based on simulation software can improve prediction accuracy, but it requires professional knowledge to guide the experiment, and the cost is very high. In this work, based on the massive geographical data and the path loss data within the service range of the base station including real data and simulation data, our approach uses the deep transfer learning method to construct a data-driven model for predicting the signal path loss of the base station. Moreover, the polar coordinate transformation is used to enhance the capacity of extracting the features of radio propagation, and a two-stream model is constructed to decouple geographic information and physical parameters of the base station for improving prediction accuracy. Using the prediction accuracy obtained between the measured and the model outputs as a measure of performance, the experimental results show that compared with the baseline model, our method significantly improves the accuracy of predicting signal path loss on testing data, which verifies the effectiveness of our method.

GIS-based Digital Simulation of Mobile Communication

A Novel Terminal-Processing Application for Utilizing Satellite Imagery in Mobile GIS

The Research of Coverage Simulation Modeling of Wireless Network Based on GIS

Research on Simulation and Test of Spatial Isolation Based on Geographic Information

Statistical Analysis of Radio Propagation Channel in Ruins Environment

A Machine Learning Based Channel Parameter Prediction Method Utilizing Ray Tracing Simulation Dataset

Modeling and Simulation on the Land Mobile Satellite Channel in Shadowing Environments

A Non-Cooperative Method for Path Loss Estimation in Femtocell Networks

Investigation of the attenuation characteristics of radio signal in ruins

A Big Data-Driven Deep Transfer Learning Approach for Path Loss Prediction in Mobile Communications.

Application Research of Multi-site Smart Substation Based on Digital and Physical Simulation Technology.

Ray-tracing propagation prediction model for mobile digital television

Modeling and simulation of geographic electromagnetic environment based on LiDAR point cloud

Path Loss Prediction for Mobile Digital TV Propagation under Viaduct

Study on the Simulation Technique for Urban Electromagnetic Environment Supported by GIS

Realtime Estimation of IEEE 802.11p for Mobile working Machines Communication respecting Delay and Packet Loss

Wireless Channel Simulation Platform for Industrial Environments Based on 3D Ray Tracing

An Empirical Analysis of the Radio Propagation Characteristics in High-Speed Railway Environment

A Radio Wave Propagation Modeling Method Based on High-Precision 3-D Mapping in Urban Scenarios

Modeling and Simulation on the Non-Frequency-Selective Land Mobile Satellite Channel in Shadowing Environments

Application of Interpolation-Based Terrain Reconstruction in Propagation Loss Prediction of Stratospheric Communications